JPH0693887B2 - Phase correction of magnetic resonance angiogram with complex difference processing - Google Patents

Description

Description: BACKGROUND OF THE INVENTION The field of the invention is nuclear magnetic resonance imaging methods and apparatus.
More specifically, the present invention provides a magnetic resonance angiogram generated using the complex difference between image data sets acquired using different values of the first moment of the gradient field. Regarding

Every nucleus that possesses a magnetic moment tries to align with the direction of the magnetic field in which it is placed. However, the nucleus then precesses around this direction at a natural nuclear frequency (Larmor frequency) related to the strength of the magnetic field and the nature of the particular nuclear species (magnetic rotation constant γ of the nucleus). The nucleus that exhibits this phenomenon is called “spin” in this specification.

When a substance such as human tissue is subjected to a uniform magnetic field (polarizing magnetic field B ₀ ), the individual magnetic moments of spins in the tissue try to align with this polarizing magnetic field, but it is irregular at the intrinsic Larmor frequency. Precess around it. A net magnetic moment M ₂ is generated in the direction of the polarization field, but magnetic components with irregular orientations in the vertical or transverse planes (xy plane) cancel each other out. However, when this substance or tissue is in the xy plane and is subjected to the action of a magnetic field close to the Larmor frequency (excitation magnetic field B ₁ ), the net uniform moment M ₂ is rotated or the xy plane is moved. "Tilt" to x
It is possible to generate a net lateral magnetic moment Mt that rotates or swivels at the Larmor frequency in the -y plane. The extent to which the net magnetic moment M ₂ is tilted (repelling angle or flip angle), and the magnitude of the net lateral magnetic moment Mt, is mainly the magnitude of the applied excitation magnetic field B ₁ and the length of time. Related to Sato.

The practical value of this phenomenon lies in the signal emitted by the excited spins after the excitation signal B ₁ ends. In a simple device, the excited spins induce an oscillating sinusoidal signal in the receiving coil. The frequency of this signal is the Larmor frequency, and its initial amplitude A ₀ is determined by the magnitude of the lateral magnetic moment Mt and the repelling angle. The amplitude A of the emitted signal decays exponentially with time t.

Decay constant 1 / T ^* ₂ is the homogeneity of the magnetic field, - related to the T _2, called the "spin-spin relaxation" constant, or "lateral relaxation" constant. The constant T ₂ is equal to the exciting magnetic field B ₁ in a completely homogeneous magnetic field.
After is removed, the precession with uniform spin is inversely proportional to the exponential velocity which is out of phase.

Another important factor contributing to the amplitude A of the NMR signal is spin-
Called the lattice relaxation process, it is characterized by the time constant T ₁ . This represents that the net magnetic moment M recovers to an equilibrium value along the magnetic polarization axis (z). The time constant T ₁ is longer than T ₂ , and much longer for most medically interesting substances.

The NMR measurements that are sometimes relevant to this invention are called "pulsed NMR measurements". This NMR measurement is divided into an RF excitation period and a signal emission period. These measurements are performed cyclically and the NMR measurements are repeated many times to accumulate different data during each cycle or to make the same measurement at different locations within the subject. A wide range of different pre-excitation schemes are known which apply one or more RF excitation pulses (B ₁ ) of varying magnitude, duration and direction. These excitation pulses may have a narrow frequency spectrum (selective excitation pulse), or a broad frequency spectrum (non-selective excitation pulse) that produces lateral magnetization Mt over a range of resonant frequencies. You may have it.
Conventionally, it has been designed to utilize a specific NMR phenomenon.
There are many excitation methods that solve certain problems in the MR measurement process.

When making an image using NMR, a method of obtaining an NMR signal from a specific place in the subject is used. Typically, the area to be imaged (the area of interest) is scanned by a series of NMR measurement cycles that vary depending on the particular localization method used. The resulting set of received NMR signals is digitized and processed to reconstruct the image using one of many well known reconstruction methods. In order to perform such a scan, of course,
It is necessary to extract the NMR signal. This is the polarization field B
It is achieved by using magnetic fields (Gx, Dy, Gz) which have the same direction as ₀ , but have gradients along the x, y and z axes respectively. By controlling the strength of these gradients during each NMR cycle, the spatial distribution of spin excitations can be controlled and the location of the resulting NMR signal can be localized.

The NMR data for reconstructing the image can be collected using one of many methods available. Typically,
These methods use a pulse sequence consisting of a plurality of diagrams that are sequentially constructed. Each figure has one or more
NMR experiments can be included, each experiment being at least RF
It has an excitation pulse and a magnetic field gradient pulse for encoding spatial information in the resulting NMR signal. As is known, the NMR signal may be a free induction decay (FID) or spin echo signal.

The present invention will be described in detail with respect to a modification of a well-known Fourier transform (FT) image forming method generally called "spin-warp type". The spin twist method is described in Physics in Medicine and Biology, Vol. 25,
WA Edelstein et al., Pp. 751 to 756 (1980), "Spin-Twist NMR Imaging and Its Application to Human Whole-Body Imaging." This is the NMR spin
Prior to collecting the echo signals, a variable amplitude phase-encoded magnetic field gradient is used to encode the spatial information in phase in the direction of this gradient. For example, in the case of a two-dimensional type (2DFT), spatial information in this direction is encoded by applying a phase encoding gradient (Gy) along one direction, and then a direction orthogonal to the phase encoding direction. The spin echo signal is collected in the presence of the read magnetic field gradient (Gx) at. The readout gradient present during the acquisition of the spin echo signal encodes the spatial information in the orthogonal direction. In a typical 2DFT pulse sequence, the magnitude of the phase-encoded gradient pulse Gy is incremented (ΔGy) in the series of views acquired during the scan, from which the NMR data set is reconstructed so that the entire image can be reconstructed. To occur.

There are a number of well-known NMR methods for measuring spin motion or flow in the region of interest. There is a "time of flight" method that excites this mass of spins as they pass a certain upstream location and inspects the resulting state of transverse magnetization at the downstream location. Determine the speed of lumps. This method has been used for many years to measure flow in tubes and more recently it has been used to measure blood flow in human limbs. Examples of this method are U.S. Pat.Nos. 3,559,044, 3,191,119, and 3,41.
No. 9,793 and No. 4,777,957.

The second flow measurement method is the inflow / outflow method, in which the spins in one localized volume or slice are excited, and after a short time the resulting change in transverse magnetization is examined. The effect of excited spins flowing out of this volume or slice, as well as the effects of differently excited spins flowing into this volume or slice are measured. Examples of this method are U.S. Pat.Nos. 4,574,239 and 5,5,5.
32,474 and 4,516,582.

A third method of measuring motion or flow relies on the fact that the NMR signals generated by spins flowing in a magnetic field gradient undergo a phase proportional to velocity. This is called the "phase modulation" method in the industry. For a flow with a roughly constant velocity during the measurement cycle, the phase change of the NMR signal is expressed as:

Δφ = γ M ₁ v Here, M ₁ is the first moment of the magnetic field gradient, γ is the gyromagnetic ratio, and v is the spin velocity. To eliminate the error introduced into this measurement due to phase shifts due to other causes, U.S. Pat.
It is common to make this measurement at least twice with different magnetic field gradient moments as described in 872. The phase difference occurring anywhere between the two measurements is then:

Δφ = γΔM ₁ v By performing two full scans using the first moments of different magnetic field gradients and subtracting the measured phase of the reconstructed image at each location in the collected data array, A phase map is created that accurately measures the speed of constantly moving spins.

A magnetic resonance angiogram is created by collecting and calculating the phase difference between at least two NMR data sets, both using different values for the first moment of the magnetic field gradient. This phase is substantially the same in the two datasets where the spin is immobile, and in the reconstructed image these tissues appear dark. On the other hand, moving spins impart a phase to the collected NMR signal that is proportional to the values of the velocity and the first moment of the magnetic field gradient. Since the two collected NMR data sets have different first moments, the moving spins cause a phase difference, and the place where these phases occur appears bright in the angiographic image.

There are numerous methods currently in use to generate phase contrast angiography data from two or more collected NMR data sets. One method, called the "Phase Difference Method", calculates the phase difference between corresponding elements in the acquired NMR image and uses this phase difference to determine the pixel of the corresponding image. Control strength. The second method, called the "complex difference method", uses the collected 2
The in-phase and quadrature components of the corresponding elements in one NMR raw data set are subtracted to produce complex difference data, which is then Fourier transformed into the image domain. Alternatively, but equivalently, the two raw NMR data sets may be Fourier transformed and the in-phase and quadrature components of the corresponding elements of the image subtracted. In any case, the complex difference image is used to reconstruct the angiographic image. There are subtle differences between angiograms made by the phase-difference method and the complex-difference method that both have unique utility in clinical applications.

Accurate reproduction of magnetic resonance angiograms using any of these phase contrast methods is premised on the absence of phase error in the measurements. There are many sources of phase error, but most phase errors are the same in the two data sets collected and are subtracted from the image data and disappear by the difference process. However, some phase errors may be different in the two data sets collected, and creating an angiogram using the difference process produces artifacts in the reconstructed image.

Numerous methods are known for correcting the phase error of NMR measurements. For example, the paper entitled "Automatic Phase Correction Method for Nuclear Magnetic Resonance Imaging" by Yuan Liu et al. In Journal of Magnetic Resonance, Vol. 86, p. 593 to p. 604 (1990). Includes (a) misalignment of the reference phase with respect to the receiver's phase detector, (b) the phase caused by the noise filter, (c) inaccurate matching of the data acquisition window, and (d) incomplete selectivity. A phase correction method is described which reduces errors due to device hardware problems such as RF excitation pulses, (e) amplifier dead time and (f) eddy currents. IEEE Transactions on Medical Imaging, M1-6, Issue 1, March 1987.
CB Arn et al., "A New Phase Correction Method for NMR Imaging Based on Autocorrelation and Histogram Analysis," and Medical Physics, Vol. 16, No. 5, September 1989.
In the January issue of MA Bern et al., "Improved Detection Capability in Magnetic Resonance Images With Low Signal-to-Noise Ratio by Phase-Corrected Actual Reproduction" describes other methods of phase correction. ing. None of these methods have been used to correct angiograms made by the complex difference method.

SUMMARY OF THE INVENTION The present invention is directed to a method and apparatus for making NMR angiograms using the complex difference method of image regions to reduce artifacts in images in such angiograms due to phase errors in the apparatus. More specifically, the present invention uses a pulse sequence having first moments of different magnetic field gradients to provide a plurality of k-space NMs.
R data sets are collected, and the plurality of k-space data sets are Fourier transformed to form a plurality of corresponding image space data sets, and a vector represented by each element in one image space data set and , Calculate the angle between each corresponding element in another image space data set and the vector it represents, store these angles as the corresponding elements in the phase difference array, and Value of the corrected phase difference array is calculated from the corresponding values, and using the corresponding value in each of the plurality of image space data sets and the corresponding value from the corrected phase difference array, the angiography It includes the step of calculating the values of the difference data set that make up the plot.

A general object of this invention is to correct for phase errors in an angiographic NMR data set generated using the complex difference method. A number of methods are known for calculating the value of the phase error from the phase difference NMR data. The present invention makes it possible to calculate such a phase error, and provides a method for correcting the image area complex difference data used for making an angiogram based on the calculation of the phase error.

The above and other objects and advantages of the present invention will be apparent from the following description. In the following description, preferred embodiments of the present invention will be described with reference to the drawings. However, the embodiments do not necessarily represent the entire scope of the present invention, and the scope of the present invention should be construed according to the claims. See range.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of an NMR apparatus using the present invention.

FIG. 2 is a block diagram of a transceiver forming part of the NMR apparatus of FIG.

3A-3C are graphs of the NMR pulse sequence used in the preferred embodiment to collect the NMR data sets used to reconstruct the angiogram according to the present invention.

FIG. 4 is a flow chart showing how an angiogram is reproduced using the NMR data set acquired by the apparatus of FIG. 1 using the pulse sequence of FIG.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Referring first to FIG. 1, the major components of a preferred NMR apparatus using the present invention are shown in block diagram form, which is a product of General Electric Company under the tradename "SIGNA". It is sold.
The overall operation of the device is controlled generally by the host computer system shown at 100, which includes a main computer 101 (such as data General M _V 7800). The computer is provided with an interface 102, through which a plurality of computer peripheral devices and other components of the NMR apparatus are coupled. Among the computer peripherals is the magnetic tape drive 104,
This can be used under the direction of the main computer to record patient data and images on tape. The processed patient data may also be stored in the image disc storage device shown at 110. The function of the image processor 108 is
Manipulating interactive image displays such as image comparison, grayscale adjustment and real-time data display. The computer system comprises means for storing raw data (ie, before reproducing the image) utilizing the disk data storage device shown at 112. An operator console 116 is also coupled to the computer via interface 102 to provide data regarding patient exams, calibration,
It provides the operator with a means to enter other data necessary for proper operation of the NMR apparatus, such as the start and end of a scan. The operator console is also used to display images stored on disk or magnetic tape.

Computer system 100 exerts control actions on the NMR system via system controller 118 and gradient amplifier device 128. Computer 100 communicates with system controller 118 via link 103 in well-known fashion. The system controller 118 includes a pulse control module (PCM) 120, an array processor 106, a radio frequency transceiver 122, a status and control module (SCM) 124, and a power supply generally indicated at 126 for energizing the components. Including some parts. PCM 1
20 utilizes a control signal supplied from the main computer 101 to control the excitation of the gradient coil as well as a digital waveform such as the RF envelope waveform used by the transceiver 122 to modulate the RF excitation pulse. Generates timing and control signals. Gradient waveform is generally Gx, Gy
And a Gz amplifier 130, 132, 134 comprising a gradient amplifier device 1
Applied to 28. Each amplifier 130, 132, 134 is utilized to excite a corresponding gradient coil in an assembly generally indicated at 136. When energized, the gradient coil produces a magnetic field gradient Gx, Gy, Gz of the magnetic field in the same direction as the main polarization field. The gradient is x-, y-, which are orthogonal to each other in the Cartesian coordinate system.
And towards the z-axis. That is, the magnetic field generated by the main magnet (not shown in the drawing) is oriented in the z-axis direction, which is called B _0, and the total magnetic field in the z-axis direction is called Bz. , Gx ＝ ∂Bz / ∂x, Gy ＝ ∂B
z / ∂y, Gz = ∂Bz / ∂z, and the magnetic field at an arbitrary point (x, y, z) is represented by B (x, y, z) = B ₀ + GxX + GyY + GzZ.

The gradient magnetic field is used in combination with the radio frequency pulses generated by the transceiver 122, the RF amplifier 123 and the RF coil 138 to encode spatial information into the NMR signal emanating from the area of the patient under examination. Pulse control module
The waveforms and control signals generated by 120 are utilized by transceiver device 122 to control the modulation and mode of the RF carrier. In transmit mode, the transmitter provides a radio frequency waveform modulated according to a control signal to an RF power amplifier 123, which is located in the main magnet assembly 146.
Energize 38. The NMR signal emitted from the excited nuclei in the patient is the same or different RF as the one used for the transmission.
It is sensed by the coil and amplified by the preamplifier 139. At the receiver portion of the transceiver 122, the NMR signal is amplified, demodulated, filtered and digitized. Dedicated unidirectional link for processed NMR signals 1
Via 05, it is sent to the array processor 106 for processing.

The PCM 120 and SCM 124 are independent devices, both in communication with peripherals of the main computer 101, such as the patient positioner 152, and also with each other via a serial communication link 103. . Each of the PCM 120 and the SCM 124 is composed of a 16-bit microprocessor (for example, Intel 80286) that processes a command from the main computer 101. The SCM 124 includes means for gathering information regarding the position of the patient cradle as well as the position of the movable patient alignment light fan beam (not shown). This information on the main computer
Used at 101 to modify image display and playback parameters. The SCM 124 also initiates functions such as patient transport and actuation of alignment devices.

Gradient coil assembly 136 and RF transmit and receive coil 138
It is mounted in the bore of the magnet used to generate the polarizing magnetic field. The magnet is part of the main magnet assembly that includes the patient alignment device 148. A shim power supply 140 is utilized to energize the shim coil attached to the main magnet. These shim coils are used to correct inhomogeneities in the polarization field. In the case of a resistive magnet, the main magnet power supply 142 is used to continuously energize the magnet. In the case of a superconducting magnet, the main power supply 142 is used to bring the polarizing magnetic field generated by the magnet to the proper operating strength, and then disconnected. In the case of a permanent magnet, the power supply 142 is not needed. A patient alignment device 148 operates in combination with a patient rocker and transport device 150 and a patient positioner device 152. Includes main magnet assembly, gradient coil assembly, RF transmit and receive coils, and patient handling equipment to minimize interference from external sources
The components of the NMR apparatus are enclosed in an RF shielded room, generally indicated at 144.

Referring particularly to FIGS. 1 and 2, the transceiver 122 causes the coil 138A to pass the RF excitation magnetic field through the power amplifier 123.
It includes a component that produces B ₁ and a component that receives the resulting NMR signal induced in coil 138B. The fundamental or carrier frequency of the RF excitation field is generated under the control of frequency synthesizer 200. The combiner receives a set of digital signals from the main computer 101 via a communication link 103. These digital signals are generated at output 201 R
F Indicates the frequency and phase of the carrier signal. RF mandated
The carrier wave is applied to modulator 202, where it is modulated in response to a signal R (t) received from PCM 120 via bus 103. The signal R (t) defines the envelope and thus the bandwidth of the RF excitation pulse to be generated. When an RF excitation pulse representing the desired envelope is generated, it is sequentially recalled by a series of stored digital values that cause it to PCM.
Raised at 120. These stored digital values can be altered by computer 100 to enable the envelope of any desired RF pulse to be generated. The magnitude of the RF excitation pulse output from line 205 is attenuated by the transmit attenuator circuit 206. This circuit 206 receives the digital signal TA from the main computer 101 via the communication link 103. The attenuated RF excitation pulse is applied to the power amplifier 123 that drives the RF transmitter coil 138A. For more information on this portion of transceiver 122, see US Pat. No. 4,952,877.

Continuing with the description of FIGS. 1 and 2, the NMR signal generated by the subject is picked up by the receiving coil 138B and applied to the input of the receiver 207. Receiver 207 amplifies the NMR signal, which is attenuated by an amount determined by the digital attenuation signal (RA) received from main computer 101 via link 103. The receiver 207 is turned on and off by a signal sent over line 211 from the PCM 120 so that the NMR signal is collected only for the duration required for the particular acquisition being performed.

The received NMR signal is at or near the Larmor frequency, which in the preferred embodiment is about 63.86 MHz. This high frequency signal is demodulated by demodulator 208 in a two step process. The demodulator first mixes the NMR signal with the carrier signal on line 201 and then mixes the resulting signal with the 2.5 MHz reference signal on line 204. The demodulated NMR signal thus obtained on line 212 has a bandwidth of 125 kHz,
It is centered around the 187.5kHz frequency. Demodulated
The NMR signal is applied to the input of analog to digital (A / D) converter 209. The converter samples and digitizes this analog signal at a rate of 250 kHz. A / D converter 2
The output of 09 is applied to the digital quadrature detector 210. This detector corresponds to the received digital signal
It produces a 16-bit in-phase (I) value and a 16-bit quadrature (Q) value. The stream of digitized I and Q values thus obtained from the received NMR signal is output to the array processor via bus 105, where it is used to reconstruct the image.

Both the modulator 202 at the transmitter portion and the demodulator 208 at the receiver portion are operated with a common signal to preserve the phase information contained in the received NMR signal. Specifically, the carrier signal at output 201 of frequency synthesizer 200 and the 2.5 MHz reference signal at output 204 of reference frequency generator 203 are used in both the modulation and demodulation processes. Thus, phase coherency is maintained so that the phase change in the received and demodulated NMR signal is an accurate representation of the phase change produced by the excited spins. 2.5MHz reference signal and 5,10,60MHz
2 are generated by a reference frequency generator 203 from a common 10 MHz clock signal, and the three reference signals described below are used in frequency synthesizer 200 to generate a carrier signal at output 201. For more information on receivers, see US Pat. No. 4,992,736.

The NMR apparatus of FIG. 1 implements a series of pulse sequences in order to collect sufficient NMR data to reproduce the desired angiogram. Referring specifically to FIG. 3A, the first pulse sequence is a normal first-order-moment-gradient-echo sequence, with the selective RF excitation pulse 300 in the presence of the G ₂ slice selective gradient pulse 301. To the subject. The excitation pulse 300 has a repelling angle α, and a typical value of α is 30
It is ゜. Compensate for FID for phase shifts by slice-selective gradient pulses 301 and F for velocity along the z-axis
In order to prevent the ID from being affected, by the G ₂ gradient coil,
A negative Gz gradient pulse 304 is followed by a positive G ₂ gradient pulse 305. For example, one solution uses a pulse 304 that is the same width as pulse 301 but has the opposite sign
05 is half the width of pulse 301, but at the same height. Although the pulses 304, 305 compensate for velocity along the z-axis, more complex gradient waveforms are well known to those skilled in the art to compensate for acceleration as well as higher order motion.

In order to encode the position of the NMR signal 303, the phase-encoded Gy gradient pulse 306 is applied to the subject shortly after the RF excitation pulse 300 is applied. As is well known, a complete scan consists of a series of these pulse sequences, in which the values of the Gy phase-encoded pulse are preserved over a series of, for example 256 discrete phase-encoded values. Along the y-axis,
Locate the spin that produces the NMR signal. The position along the x-axis is located by the Gx gradient pulse 307. This pulse is generated when collecting the NMR gradient echo signal 303, which encodes the NMR signal 303 in frequency. Unlike the Gy phase encoded gradient pulse 306, the Gx readout gradient pulse 307
Stays constant for the entire scan. Gradient echo 30
Before the pulse 307, a gradient pulse 308,3 is generated in order to generate 3 and make it insensitive to velocity along the x-axis.
Generates 09. There are many known methods of accomplishing this.

As will be described in more detail below, at least two complete data sets, each having different flow sensitivities along one direction, are required to practice the invention. In the preferred embodiment, this set of data is collected staggered. In this scheme, two or more measurements with different flow sensitivities are collected with a single value of the phase encoding gradient. The phase encoding value is then changed and an additional measurement with two or more flow sensitivities is made with this new phase encoding value. Until all phase encoding values are used,
This process continues. After that, the collected data is
Reorder the k-space NMR data sets so that they all have different flow sensitivities. This staggering scheme is preferred in that it minimizes the effects of other movements (eg, breathing), but in the following description of the invention, the k-space NMR data set will be acquired before the next flow encoding is used. Both will be described as being completely collected.

The NMR signal 303 is collected by the instrument's transceiver 122 and digitized into a row of 256 complex numbers,
These numbers are stored in the memory of the main computer 101. For each value of the Gy phase encoding gradient, the NMR signal 303 is generated, collected, digitized and stored in a separate row of 256 complex numbers. Therefore, when the scan is completed, the computer 101 has a two-dimensional complex number (256 × 25
The matrix of 6) is stored. If such an NMR signal is generated when no flow-sensitive gradient is applied, the k-space NMR data set can be Fourier transformed into a normal NMR image. This is represented in FIG. 3A by the gradient G _M. This gradient G _M is zero when the first moment of the magnetic field gradient that encodes velocity in the pulse sequence is not used.

Collected to generate an angiogram according to the present invention
The NMR signal is made sensitive to velocity. The pulse sequence used is the same as that shown in Figure 3A, but this time the gradient G _M is collected for velocities along the direction of G _M.
The difference is that it has a value that makes the NMR signal sensitive. This is shown in Figure 3B. In this figure, G _M is
A bipolar waveform composed of a negative gradient pulse 310 followed by a positive gradient pulse 311. The area (A) defined by each pulse 310, 311 is the same and each gradient pulse
The centers of 310 and 311 are separated from each other by a distance (t). Therefore, the incremental first moment (ΔM ₁ ) caused by the gradient G _M is ΔM ₁ = A × t, and the first moment ΔM ₁ of this gradient is the signal 303 after the RF excitation pulse 300 is applied.
Applied before collecting. Although the gradient G _M is shown as a separate gradient field, it is actually produced by the same coil that produces the Gx, Gy, Gz gradient fields. By combining Gx, Gy, and Gz gradient magnetic fields having appropriate magnitudes, the gradient moment G _M can be directed in any direction in space and made sensitive to the flow in that direction. For example, it is quite common to make the flow sensitive to the slice selection direction, in which case the gradient moment G _M is generated solely by the Gz gradient coil. Therefore, in the preferred embodiment, the
Using the pulse sequence of Figure 3A, but adding the _GM gradient field of Figure 3B, collect the _first array Z ₁ of the NMR raw signal.

After collecting and storing a first sequence Z ₁ of the NMR signal, it collects the second sequence Z ₂ signal. This uses the pulse sequence of FIG. 3A, but the gradient moment G _M is performed during a scan modified as shown in FIG. 3C to produce a moment of -ΔM ₁ . This is accomplished by having the gradient pulses 312, 313 be identical, but in the opposite direction as the gradient pulses 310, 311. 256 NMR signals Z ₂ were collected,
After being stored in the computer 101, the data acquisition phase for one axis of motion is complete and the data processing phase begins.

It will be apparent to those skilled in the art that various modifications can be made in the data collection stage of the present invention. Other NMR pulse sequences may be used and, as previously mentioned, k-space NM
The collection of R data sets Z ₁ and Z ₂ can be different from each other. Further, as described in U.S. Pat. No. 4,443,760, multiple orders may be implemented with each phase encoding gradient Gy to cancel out instrumental errors or improve the signal to noise ratio. . There are many different ways to generate the gradient moment ΔM ₁ using the gradient G _M. For example, the gradient pulses 310-313 may have different shapes, or they may be separated in time to increase the _first moment ΔM ₁ . It is also possible to use a spin-echo sequence using 180 ° RF pulses to refocus on the unwanted effects of static magnetic field inhomogeneities. When using 180 ° pulses, it is well known that the first moment can be generated by placing gradient lobes of the same polarity on either side of the 180 ° RF pulse.

In a preferred embodiment of this invention, three pairs of k-space NMR data sets were collected as described above. First pair Z ₁
(X) and Z ₂ (x) have a first moment of gradient G _M oriented towards the x-axis and the second pair Z ₁ (y) and Z ₂ (y) are
has a first moment of gradient G _M directed toward the y-axis,
The pair Z ₁ (z) and Z ₂ (z) of has the first moment G _M of the gradient pointing towards the z-axis. As will be explained later, these pairs of k-space NMR data sets are processed separately according to the invention and the results combined to form a spin-sensitive angiogram in any direction.

Although the six-point method of measuring spin velocity is described above, the present invention is also compatible with a simple four-point method of measuring three-dimensional flow. In this case, three data sets Z ₂ (x), Z ₂
Replace (y) and Z ₂ (z) with one data set Z ₂ (ref) generated using a flow-compensated pulse sequence that is unaffected by movement along any axis. Both the 6-point method and the balanced 4-point method have been published by the inventor in the Society of Magnetic Resonance in Medicine, Abstract No. 475 (1990) of the 9th Annual General Assembly "for phase contrast flow measurement. Optimal Coding of. "

The processing of the k-space NMR data sets Z ₁ and Z ₂ to produce an angiogram is illustrated in FIG. All the processes are executed by the main computer 101 under the instruction of the instructions in the storage program. Two k-space NMR data sets Z ₁ and Z ₂ are the complex numbers 256 × 256 shown in blocks 320 and 321.
It is stored as an array. The first step in the method is to perform a two-dimensional complex Fourier transform on each of the data sets 320, 321 to transform the images they represent from k-space to the image domain. This is the same transformation that is used to generate a normal NMR image, the result of which is a two 256 × 256 element array 322,323 of complex numbers (M ₁ ) and (M ₂ ).
Is. The magnitudes of the complex numbers M ₁ (x, y) and M ₂ (x, y) indicate the spin densities at the corresponding pixel positions (x, y) of the image, and the respective complex numbers M ₁ (x, y) and M _{The 2} (x, y) phase is determined by both the velocity of the spin along the direction of the magnetic field gradient G _M that encodes the motion and the phase error of the device, if any. It is the phase error of these devices that produces the image artifacts when generating a normal complex difference angiogram using the NMR data sets M ₁ and M ₂ in the image domain, and using the idea of the invention. It is these phase errors that are substantially reduced by.

Continuing with the explanation of FIG. 4, the phase correction value is the NMR of the image region.
Calculated from the data sets 322,323. To be specific,
256 × 25 from two transformed NMR data sets M ₁ and M ₂
A 6-element phase difference array 325 is calculated. Each element (φ) of the phase difference array 325 is a respective data set as shown below.
Calculated from the corresponding I ₁ , Q ₁ and I ₂ , Q ₂ elements of 322,323.

φ = tan ^-1 Q ₁ / I ₁ −tan ^-1 Q ₂ / I ₂ ; or φ = tan ^-1 [(I ₂ Q ₁ −I ₁ Q ₂ ) / (I ₁ I ₂ + Q ₁ Q ₂ )] Here M ₁ = I ₁ + iQ ₁ | M ₁ | = (I ₁ ² + Q ₁ ² ) ^1/2 M ₂ = I ₂ + iQ ₂ | M ₂ | = (I ₂ ² + Q ₂ ² ) ^1/2 top The value of φ shown in is preferably calculated using the four quadrant arctangent.

After correcting each element (φ) of the phase difference array 325,
Generate the corresponding element (φcor) of the 56 × 256 element corrected phase difference array 326. This phase correction value can be represented by a model called a sum of a constant component and two linear components.

_{φcor = φ- (φ 0 + αx} + βy) [1] Thisゝx and y is the position of the two coordinates of the imaging. The method described here is general enough to be compatible with almost any phase correction method, but in the preferred embodiment φ
₀ , α and β are calculated using a least squares fit.
It has been found that weighting the least squares fit by the square of the average magnitude (A = (| M ₁ | + | M ₂ |) / 2) adequately suppresses small signal regions. Therefore, the fit of least squares to φ ₀ , α, and β is as follows.

The sum of this equation spans the values of all pixels in the image array.
Thus, using the above formula for φ cor, and calculating φ ₀ , α and β using the average magnitude value (A) and the phase difference value from the array 325, for each position x, y And the corrected phase φcor is calculated. One important advantage of making a phase correction on a phase image is that it is easy to extend this correction to higher orders than the first order.

The complex difference NMR data set is then calculated using the image region NMR data sets 322, 323 and the corrected phase difference array 326. This is the corresponding element M ₁ , M ₂ of the arrays 322, 323, 326, respectively.
And φcor and the following formula, the complex difference data set 3
This is accomplished by computing each of the 30 256 × 256 elements (D).

Dx = (| M ₁ | ² + | M ₂ | ² −2 | M ₁ || M ₂ | cosφcor) ^{1/2 In} other words, the size data in the NMR data set 322,323 in the image region is used. Do not use the phase difference information. Instead, the corrected phase difference data φcor in array 326 is used to eliminate artifacts that cause phase errors from the resulting angiogram.

The procedure described above is repeated for each motion axis (x, y, z) to generate two more 256 × 256 element complex difference NMR data sets 331,332. Thus, the difference information in arrays 330-332 indicates the complex difference Dx, Dy, Dz due to the motion of the spins along the x, y and z axes, respectively. These are combined to generate a 256 × 256 element angiogram NMR data set 335 according to the following equation:

D = ((Dx) ² + (Dy) ² + (Dz) ² ) ^{1/2 Using} this NMR data set 335 for the angiogram, a display in which moving spins are brighter than immobile spins is generated. You can do it. Since the motion of blood through the cardiovascular system is the main motion, the angiogram is essentially an image of the cardiovascular tree in view.

Continuation of front page (51) Int.Cl. ⁵ Identification number Office reference number FI Technical indication location G01R 33/28 8825-4C A61B 5/05 312 9219-2J G01N 24/02 Y (56) References 4-288143 (JP, A) JP-A-2-200247 (JP, A) JP-A-63-40540 (JP, A)

Claims (3)

[Claims] 1. An NMR apparatus for producing an angiogram with reduced sensitivity to unwanted phase shifts, means for applying a polarizing magnetic field to a region of interest and a magnetic field gradient to the region of interest. Means for applying an RF excitation field to the region of interest, means for collecting NMR signals generated by excited spins within the region of interest, and using a first motion-encoded magnetic field gradient First NMR signal collected with
Means for performing a first scan, which is stored as a k-space NMR data set (Z ₁ ) of the second, and the acquired NMR signal using the second motion-encoded magnetic field gradient and a second k-space NMR data set. Means for performing a second scan, stored as (Z ₂ ),
A means for generating a corrected complex difference NMR data set from the k-space NMR data sets (Z ₁ ) and (Z ₂ ), such that the brightness of the pixels therein is determined by the values in the complex difference NMR data set. NMR apparatus having means for generating an image. 2. Means for generating a corrected complex difference NMR data set (M ₁ ) by performing a Fourier transform on the first k-space NMR data set (Z ₁ ) to obtain a NMR data set (M) of the first image region. Means for generating ₁ ) and the second k-space NMR
A means for generating a _second image domain NMR data set (M ₂ ) by performing a Fourier transform on the data set (Z ₂ ), and a first and second image domain NMR data set (M ₁ ) and Means for calculating the phase difference array (φ) for calculating the value of each element therein from the value of the corresponding element in (M ₂ ),
Means for calculating a corrected phase difference array (φcor) such that the value of each element therein is calculated from the value of the corresponding element in said phase difference array (φ); The element (D) is the NMR data set M of the first and second image regions.
₁ and M ₂ and the corrected phase difference array (φcor) are calculated according to the following equation D = (| M ₁ | ² + | M ₂ | ² −2 | M ₁ || M ₂ | cos φcor) ^1/2 The NMR apparatus according to claim 1, further comprising means for calculating a complex number difference NMR data set. 3. The first and second motion-encoded magnetic field gradients are applied separately along each of the three axes of motion, and the NMR apparatus further comprises a complex difference number NMR data set for each of the three axes of motion ( Dx), (Dy) and (Dz), and the means for generating the image are data from corresponding elements of three complex difference NMR data sets (Dx), (Dy) and (Dz). Combining to determine the brightness of the pixels of the image.
The described NMR apparatus.